Desktop videoconferencing (DVC) offers many opportunities for language learning through its multimodal features. However, it also brings some challenges such as gaze and mutual gaze, that is, eye-contact. This paper reports some of the findings of a PhD study investigating social presence in DVC interactions of English as a Foreign Language (EFL) teacher trainees. The case study approach involved the exploration of online interactions of five cases (pairs) within an interpretivist paradigm. Data collection included interviews, questionnaires and analysis of DVC recordings. The study emphasizes the importance of eye-contact in online multimodal communication to facilitate the establishment of social presence. Five types of gaze that were observed in learner interactions and participants' perspectives on eye-contact are reported. The conclusions include technical suggestions for the use of a webcam as well as pedagogical implications of online video interaction.